Table 2. Summary characteristics of included studies (N=84).
A-CACXL: accelerated contact lens-assisted corneal cross-linking; A-CXL: accelerated corneal cross-linking; ATE-CXL: accelerated transepithelial corneal cross-linking (a high-intensity, short-duration epi-on protocol); cl-ACXL: continuous light-accelerated corneal cross-linking; CLCXL: contact lens-assisted corneal cross-linking (a modified accelerated protocol using a contact lens to aid in treatment for thin corneas); cCXL: customized corneal cross-linking; CRXL: corneal cross-linking with mechanical compression (an accelerated protocol involving mechanical compression of the cornea using a flat rigid contact lens during CXL); CXL: corneal collagen cross-linking; EFPL-M-TECXL: transepithelial enhanced fluence pulsed light accelerated CXL; Epi-off: epithelium-off protocols (conventional Dresden method); Epi-on: epithelium-on or transepithelial protocols, preserves the corneal epithelium; Epioxa™: FDA-approved epi-on CXL using oxygen-enriched RF for keratoconus; EpiSmart: investigational epi-on CXL with RiboStat (RF + sodium iodide); eRF: enhanced riboflavin; I-CXL: transepithelial corneal collagen crosslinking using iontophoresis (riboflavin delivery via iontophoresis); NR: not reported; NRCT: non-randomized comparative trial; pl-ACXL: pulsed light-accelerated corneal cross-linking; PRK: photorefractive keratectomy; RCT: randomized controlled trial; RF: riboflavin (the photosensitizer used in CXL protocols); STARE-X: selective transepithelial topography-guided photorefractive keratectomy combined with simultaneous accelerated corneal crosslinking; T-CAT: topography-guided customized ablation treatment (a refractive procedure combined with CXL); TA-CXL: topo-pachimetric accelerated epi-on CXL (customized based on topography and pachymetry); TE-CXL: transepithelial corneal cross-linking (an epi-on variant preserving the epithelium)
| Protocol Group | Study | Study Design | CXL Protocol(s) | Sample Size | Follow-up Duration |
| Conventional Dresden / Standard Epi-off | Alqudah et al., 2025 [14] | Retrospective observational cohort | Dresden vs I-CXL | 107 patients | 1 year |
| Bhattacharyya et al., 2019 [15] | Prospective observational cohort | Dresden vs Control | 78 eyes | 6 months | |
| Choi et al., 2017 [16] | Prospective RCT | Dresden vs A-CXL | 28 eyes | 6 months | |
| Danesh et al., 2021 [17] | Prospective observational cohort | Dresden | 31 eyes | 1 year | |
| Dervenis et al., 2020 [18] | Retrospective case series | Dresden vs A-CXL | 59 eyes | 7 months | |
| Elghobaier et al., 2025 [19] | Retrospective case series | Dresden | 223 eyes | 3 years | |
| Elmassry et al., 2021 [20] | Retrospective observational cohort | Dresden vs TE-CXL vs A-CXL | 6120 eyes | 10 years | |
| Ferdi et al., 2019 [21] | Systematic review and meta-analysis | Epi-off (various) | 41 studies | 1 year | |
| Ferdi et al. (Registry), 2023 [2] | Observational registry study | Dresden | 162 eyes | 5 years | |
| Gustafsson et al., 2025 [22] | Prospective RCT (non-inferiority) | Iso vs Hypo-osmolar RF | 54 patients | 1 year | |
| Hallahan et al., 2014 [23] | Prospective RCT | Dresden | 51 eyes | 3 months | |
| Kandel et al. (Registry), 2024 [24] | Observational registry study | Dresden | 176 eyes | 5 years | |
| Kymionis et al., 2016 [25] | Prospective NRCT | Dresden vs cCXL | 32 eyes | 1 month | |
| Lindstrom et al., 2021 [26] | Economic modeling study | Dresden vs no CXL | Simulated cohort | Lifetime | |
| Marafon et al., 2020 [27] | Retrospective comparative cohort | Dresden vs A-CXL | 113 eyes | Mean 37.6 months | |
| Padmanabhan et al., 2014 [28] | Prospective NRCT | Dresden vs T-CAT+CXL | 27 eyes | Mean 7.7 months | |
| Price et al., 2018 [29] | Prospective interventional case series | Dresden vs A-CXL | 644 eyes | Median 3.5 years | |
| Rosenblat et al., 2014 [30] | Prospective NRCT | Standard vs hypotonic RF | 39 patients | 1 year | |
| Rosenblat et al., 2016 [31] | Prospective RCT | Standard vs hypotonic RF | 48 eyes | 1 year | |
| Seyedian et al., 2015 [32] | Prospective RCT | Dresden vs Control | 52 eyes | 1 year | |
| Soeters et al., 2014 [33] | Retrospective observational cohort | Dresden | 119 eyes | 1 year | |
| Soeters et al., 2015 [7] | Prospective RCT | Dresden vs TE-CXL | 39 eyes | 1 year | |
| Strmeňová et al., 2015 [34] | Retrospective observational cohort | Dresden | 88 eyes | 2 years | |
| Vandevenne et al., 2023 [35] | Prospective RCT (non-inferiority) | Dresden vs cCXL | 124 patients | 1 year | |
| Wittig-Silva et al., 2014 [36] | Prospective RCT | Dresden vs Control | 100 eyes | 3 years | |
| Transepithelial / Epi-on | Al Fayez et al., 2015 [37] | Prospective RCT | TE-CXL vs Dresden | 70 eyes | Mean 40 months |
| APRICITY-A, 2023 [38] | Prospective RCT | EpiSmart vs Sham | 400 subjects | 1 year | |
| APRICITY-B, 2023 [39] | Prospective RCT | EpiSmart vs Sham | 400 subjects | 1 year | |
| Bikbova et al., 2016 [40] | Prospective RCT | I-CXL vs Dresden | 149 eyes | 2 years | |
| Caruso et al., 2016 [41] | Prospective NRCT | TE-CXL | 25 eyes | 2 years | |
| Caruso et al., 2021 [42] | Prospective RCT | TA-CXL vs Dresden | NR | 2 years | |
| Cassagne et al., 2014 [43] | Prospective interventional case series | I-CXL | NR | NR | |
| Cifariello et al., 2018 [44] | Prospective RCT | TE-CXL vs Dresden | 40 eyes | 2 years | |
| Epstein et al., 2022 [45] | Prospective RCT | EpiSmart vs Control | 1922 subjects | 1 year | |
| Ferrini et al., 2023 [46] | Prospective observational cohort | TE-CXL vs Dresden | NR | NR | |
| Gatzioufas et al., 2016 [47] | Prospective observational cohort | TE-CXL | 26 eyes | 1 year | |
| Lesniak et al., 2014 [48] | Prospective RCT | TE-CXL | 30 eyes | 6 months | |
| Lombardo et al., 2016 [49] | Prospective RCT | I-CXL vs Dresden | 34 eyes | 6 months | |
| Lombardo et al., 2017 [50] | Prospective RCT | I-CXL vs Dresden | 34 eyes | 2 years | |
| Mastropasqua et al., 2014 [51] | Experimental laboratory (ex vivo) study | I-CXL vs TE-CXL vs Dresden | 10 donor corneas | Immediate (procedural) | |
| Mazzotta et al., 2020 [52] | Prospective interventional case series | ATE-CXL + O2 | 27 eyes | 6 months | |
| Mazzotta et al., 2022 [53] | Prospective NRCT | EFPL-M-TECXL + eRF | 40 eyes | 3 years | |
| Rechichi et al., 2021 [54] | Prospective NRCT | STARE-X | 100 eyes | ≥2 years | |
| Shao et al., 2020 [55] | Prospective observational cohort | TE-CXL | 46 eyes | 6 months | |
| Shetty et al., 2014 [56] | Retrospective case series | TE-CXL vs Dresden vs A-CXL (epi-off) | 2350 patients | ≥6 months | |
| Smith et al., 2025 [57] | Prospective RCT | Epioxa™ vs Sham | 312 eyes | 1 year | |
| Stojanovic et al., 2014 [58] | Prospective RCT | TE-CXL vs Epi-off | 40 eyes | Median 2 years | |
| Stulting et al., 2018 [59] | Prospective observational cohort | TE-CXL | 512 eyes | 2 years | |
| Sun et al., 2018 [60] | Prospective NRCT | TE-CXL + O2 | 26 eyes | 1 year | |
| Vinciguerra et al., 2016 [61] | Prospective NRCT | I-CXL vs Dresden | 34 eyes | 1 year | |
| Accelerated | Abdel-Radi et al., 2023 [62] | Prospective interventional case series | A-CXL | 45 eyes | 6 months |
| Adapted Fluence Study, 2015 [63] | Prospective RCT | A-CXL (7 vs 10 min) | 40 patients | 1 year | |
| Ang et al., 2025 [64] | Retrospective observational study | A-CXL | 70 eyes | 1 year | |
| Asgari et al., 2018 [65] | Prospective NRCT | A-CXL (18 vs 9 mW) | 60 eyes | 1 year | |
| Avni-Zauberman et al., 2021 [66] | Retrospective cohort | A-CXL vs Dresden | 124 eyes | 1 year | |
| Badawi, 2021 [67] | Retrospective cohort | A-CXL vs TE-CXL vs Dresden | 104 eyes | 1 year | |
| Bunin et al., 2025 [68] | Retrospective cohort | A-CXL vs A-CACXL | 62 eyes | 3 years | |
| Dina et al., 2025 [69] | NRCT (non-inferiority) | A-CXL vs Dresden | 41 eyes | 1 year | |
| Dongre et al., 2024 [70] | Retrospective case series | A-CXL vs Dresden | 964 eyes | 1 year | |
| Gupta et al., 2024 [71] | Retrospective comparative cohort | CACXL vs TE-CXL vs Dresden | 94 eyes | 1 year | |
| Hagem et al., 2017 [72] | Prospective RCT | A-CXL vs Dresden | 40 eyes | 1 year | |
| Hagem et al., 2019 [73] | Prospective RCT | A-CXL vs Dresden | 40 eyes | 2 years | |
| Hashemi et al., 2015 [74] | Prospective RCT | A-CXL vs Dresden | 62 eyes | 6 months | |
| Hashemi et al., 2017 [75] | Prospective interventional case series | A-CXL (18 vs 9 mW) | 62 eyes | 1 year | |
| Herber et al., 2018 [76] | Retrospective case series | A-CXL (beam profiles) | 45 eyes | 1 year | |
| Iqbal et al., 2019 [77] | Prospective NRCT | A-CXL+PRK vs Dresden | 125 eyes | 2 years | |
| Ishii et al., 2022 [78] | Prospective NRCT | ATE-CXL | 34 eyes | 3 years | |
| Kandel et al., 2021 [79] | Observational registry study | A-CXL vs Dresden | 684 eyes | 1 years | |
| Karotkar et al., 2022 [80] | Prospective RCT | pl-ACXL vs cl-ACXL | 100 eyes | 1 year | |
| Kobashi et al., 2020 [12] | Systematic review and meta-analysis | A-CXL vs Dresden | 6 RCTs | 1 year | |
| Kortuem et al., 2017 [81] | Retrospective comparative cohort | A-CXL vs Dresden | 286 eyes | 3 years | |
| Males et al., 2018 [82] | Retrospective comparative cohort | A-CXL vs Dresden | 42 eyes | ≥1 years | |
| Manumuraleekrishna et al., 2024 [83] | Prospective interventional case series | A-CXL (hypo- vs iso-osmolar RF) | 100 eyes | 1 year | |
| Mazzotta et al., 2021 [84] | Prospective NRCT | A-CXL | 156 eyes | 5 years | |
| Ng et al., 2016 [85] | Prospective NRCT | A-CXL vs Dresden | 26 eyes | Mean 13.9 months | |
| Ozsaygili et al., 2021 [86] | Retrospective comparative cohort | A-CXL | 64 eyes | 1 year | |
| Recalde et al., 2019 [87] | Prospective observational cohort | A-CXL | 22 eyes | 1 year | |
| Rehnman et al., 2015 [88] | Prospective RCT | CRXL vs Dresden | 60 eyes | 6 months | |
| Sherif et al., 2016 [89] | Prospective interventional case series | pl-ACXL | 20 eyes | Intraoperative | |
| Šklebar et al., 2025 [90] | Prospective observational cohort | A-CXL vs Dresden | 38 eyes | 9 months | |
| Tomita et al., 2014 [91] | Prospective NRCT | A-CXL vs Dresden | 48 eyes | 1 year | |
| Turhan et al., 2019 [92] | Prospective interventional case series | A-CXL | 52 eyes | NR | |
| Turunç et al., 2025 [93] | Prospective interventional case series | A-CXL | 113 eyes | 6 months | |
| Yousif et al., 2023 [94] | Retrospective comparative cohort | pl-ACXL vs cl-ACXL | 200 eyes | 1 year |